Table 2.
Different feature extraction algorithms using SVM, KNN, and PSO-H-ELM classifiers to classify the accuracy (%).
| Feature extraction algorithm | Classification algorithm | 1 | 2 | 3 | 4 | 5 | 6 | Average accuracy |
|---|---|---|---|---|---|---|---|---|
| Power spectral density | SVM | 81.43 | 83.59 | 94.38 | 89.76 | 70.79 | 74.13 | 83.33 |
| KNN | 77.39 | 77.57 | 97.33 | 87.57 | 88.58 | 89.32 | 86.28 | |
| PSO–H-ELM | 81.43 | 80.76 | 97.73 | 97.06 | 85.75 | 85.26 | 88.83 | |
|
| ||||||||
| EMD | SVM | 90.75 | 85.38 | 88.25 | 100.00 | 97.29 | 95.67 | 93.62 |
| KNN | 77.25 | 83.33 | 88.71 | 100.00 | 97.25 | 95.67 | 90.94 | |
| PSO–H-ELM | 89.60 | 88.85 | 94.25 | 69.57 | 99.85 | 98.23 | 94.24 | |